MiniMax · model
MiniMax M3.1 Flash benchmark scores
As of 2026-10-29, MiniMax M3.1 Flash (MiniMax) has measured scores on 21 benchmarks and estimated scores on 108 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 90.9% | estimated ± 2.7 pp, high confidence |
| AA-HLE | 48.0% | measured |
| AA-SciCode | 53.4% | measured |
| CritPt | 22.3% | measured |
| GDPval-AA | 47.9% | measured |
| GPQA Diamond | 92.9% | estimated ± 2.0 pp, high confidence |
| HLE | 53.0% | estimated ± 6.6 pp, medium confidence |
| AA-LCR | 79.7% | measured |
| LiveCodeBench v6 | 95.3% | estimated ± 1.9 pp, medium confidence |
| SciCode | 53.1% | estimated ± 3.1 pp, high confidence |
| Terminal-Bench 2.1 | 86.6% | estimated ± 5.2 pp, low confidence |
| BrowseComp | 91.6% | estimated ± 1.9 pp, high confidence |
| HLE w/ tools | 59.0% | estimated ± 4.9 pp, medium confidence |
| LongBench v2 | 62.1% | estimated ± 5.8 pp, low confidence |
| AA Coding Index | 77.3% | estimated ± 2.5 pp, high confidence |
| AA-MMMU-Pro | 79.1% | measured |
| APEX-Agents | 40.3% | estimated ± 9.0 pp, low confidence |
| APEX-Agents-AA | 38.7% | estimated ± 2.9 pp, low confidence |
| BioMysteryBench (human-difficult) | 48.6% | estimated ± 6.8 pp, low confidence |
| DeepSearchQA | 89.4% | estimated ± 9.6 pp, medium confidence |
| SWE-bench Verified | 82.0% | estimated ± 2.3 pp, high confidence |
| AutomationBench | 62.2% | measured |
| BFCL v4 | 76.5% | estimated ± 2.7 pp, medium confidence |
| CyberGym | 82.2% | measured |
| JobBench | 64.9% | estimated ± 4.9 pp, medium confidence |
| SWE-bench Pro | 62.4% | estimated ± 3.5 pp, high confidence |
| WideResearch | 83.0% | estimated ± 1.6 pp, medium confidence |
| AA Agentic Index | 42.2% | estimated ± 2.9 pp, high confidence |
| MMLU-Pro | 85.7% | estimated ± 2.5 pp, low confidence |
| Artificial Analysis Intelligence Index | 28.9% | measured |
| GPQA | 93.3% | estimated ± 3.3 pp, high confidence |
| MMLU | 94.4% | estimated ± 1.0 pp, medium confidence |
| AA-Omniscience Accuracy | 45.1% | measured |
| MMMLU | 91.0% | estimated ± 0.7 pp, low confidence |
| MMMU | 85.9% | estimated ± 2.7 pp, high confidence |
| AA MMLU-Pro | 88.3% | estimated ± 0.6 pp, high confidence |
| Gert Labs | 68.8% | estimated ± 6.0 pp, low confidence |
| AA-AnalystAgent | 48.4% | estimated ± 8.5 pp, low confidence |
| AA Harvey LAB v1.0 | 92.7% | estimated ± 1.3 pp, medium confidence |
| AA Tau3 Banking | 39.5% | estimated ± 3.3 pp, medium confidence |
| AA Terminal-Bench 2.1 | 86.2% | estimated ± 4.8 pp, medium confidence |
| ApprenticeBench | 19.4% | estimated ± 4.0 pp, medium confidence |
| ARC-AGI-1 | 94.0% | estimated ± 4.0 pp, high confidence |
| ARC-AGI-2 | 73.2% | estimated ± 13.6 pp, low confidence |
| CursorBench 3.1 | 52.4% | estimated ± 3.4 pp, medium confidence |
| CursorBench 3.2 | 59.1% | estimated ± 5.3 pp, medium confidence |
| FrontierCode 1.1 Main | 53.8% | estimated ± 2.8 pp, medium confidence |
| FrontierSWE v2 | 50.8% | estimated ± 10.4 pp, low confidence |
| MazeBench | 1.0% | estimated ± 5.4 pp, low confidence |
| OfficeQA Pro | 51.7% | estimated ± 6.8 pp, medium confidence |
| OSWorld-Verified | 79.5% | estimated ± 4.7 pp, high confidence |
| ParseBench | 65.0% | estimated ± 5.5 pp, medium confidence |
| Vals GPQA Diamond | 90.8% | estimated ± 3.0 pp, high confidence |
| Vals MMLU-Pro | 88.6% | estimated ± 1.9 pp, high confidence |
| Vals SWE-bench | 100.0% | estimated ± 4.5 pp, medium confidence |
| Terminal-Bench 2.1 (Vals AI) | 76.9% | estimated ± 5.4 pp, low confidence |
| AA AutomationBench | 72.5% | measured |
| CWE-Bench-AA | 58.6% | estimated ± 4.5 pp, medium confidence |
| AA Cyber Index | 46.3% | estimated ± 9.4 pp, low confidence |
| DeepsecBench-AA | 34.3% | estimated ± 5.3 pp, medium confidence |
| GDP.pdf | 20.8% | measured |
| AA ITBench | 48.5% | estimated ± 4.9 pp, medium confidence |
| AA Terminal-Bench 4.0 | 25.8% | measured |
| Bug Hunt Bench | 38.2% | estimated ± 6.3 pp, low confidence |
| CursorBench 4.0 | 35.6% | estimated ± 3.5 pp, medium confidence |
| CWE-bench v1 | 67.4% | estimated ± 2.7 pp, medium confidence |
| DeepSWE | 73.7% | measured |
| FrontierCode 2.0 | 29.0% | estimated ± 7.6 pp, medium confidence |
| GraphWalks BFS 256K–1M | 71.9% | estimated ± 14.8 pp, low confidence |
| HLE w/o tools | 43.4% | measured |
| OSWorld 2.0 | 61.7% | measured |
| OSWorld 2.1 | 79.8% | estimated ± 2.8 pp, medium confidence |
| PostTrainBench v1.1 | 39.3% | measured |
| ProgramBench | 90.9% | estimated ± 4.8 pp, medium confidence |
| Terminal-Bench 4.0 | 38.1% | measured |
| Toolathlon-Verified | 79.5% | estimated ± 1.0 pp, medium confidence |
| OSWorld 2.1 Offline | 55.9% | estimated ± 7.9 pp, low confidence |
| ARC-AGI-3 | 1.7% | estimated ± 7.7 pp, medium confidence |
| ARC-AGI-3 Provider Adapter | 31.5% | estimated ± 7.5 pp, medium confidence |
| Hermes Bench | 69.4% | estimated ± 2.9 pp, high confidence |
| Hermes Index | 40.2% | estimated ± 7.7 pp, medium confidence |
| CharXiv | 82.5% | measured |
| CharXiv w/o tools | 77.8% | estimated ± 2.2 pp, high confidence |
| ExploitGym | 17.0% | estimated ± 9.3 pp, low confidence |
| AI2D_TEST | 93.5% | estimated ± 0.9 pp, low confidence |
| C-Eval | 94.2% | estimated ± 1.0 pp, low confidence |
| CountBench | 97.6% | estimated ± 2.9 pp, medium confidence |
| DynaMath | 86.4% | estimated ± 0.8 pp, low confidence |
| ERQA | 68.5% | estimated ± 2.9 pp, high confidence |
| MathVision | 89.5% | estimated ± 3.2 pp, high confidence |
| MCP Atlas | 83.4% | estimated ± 10.4 pp, low confidence |
| MLVU (M-Avg) | 86.5% | estimated ± 1.5 pp, high confidence |
| MMLU-Redux | 96.6% | estimated ± 0.8 pp, low confidence |
| MMMU-Pro | 79.9% | estimated ± 1.6 pp, high confidence |
| NL2Repo | 61.2% | measured |
| QwenClawBench | 94.0% | estimated ± 6.0 pp, low confidence |
| RealWorldQA | 84.4% | estimated ± 1.5 pp, high confidence |
| ScreenSpot Pro | 76.0% | estimated ± 7.1 pp, medium confidence |
| SuperGPQA | 90.6% | estimated ± 9.3 pp, medium confidence |
| Toolathlon | 55.9% | estimated ± 6.3 pp, medium confidence |
| V* | 100.0% | estimated ± 7.5 pp, medium confidence |
| Video-MME (w/o subtitle) | 86.0% | estimated ± 1.7 pp, medium confidence |
| Video-MME (with subtitle) | 87.6% | estimated ± 2.9 pp, high confidence |
| VideoMMMU | 85.7% | estimated ± 1.5 pp, medium confidence |
| We-Math | 84.6% | estimated ± 7.9 pp, low confidence |
| HealthBench Hard | 33.0% | estimated ± 8.6 pp, low confidence |
| MedXpertQA (MM) | 76.0% | estimated ± 6.4 pp, low confidence |
| MedXpertQA (Text) | 60.8% | estimated ± 8.9 pp, low confidence |
| React Native Evals | 78.1% | estimated ± 4.7 pp, medium confidence |
| ResearchClawBench | 19.3% | estimated ± 2.2 pp, low confidence |
| Finance Agent v2 | 62.5% | estimated ± 4.5 pp, low confidence |
| BioMysteryBench (human-solvable) | 89.3% | estimated ± 0.6 pp, medium confidence |
| DRACO | 81.2% | estimated ± 3.1 pp, medium confidence |
| HealthBench (raw) | 55.7% | estimated ± 6.0 pp, medium confidence |
| HealthBench (length-adjusted) | 53.9% | estimated ± 3.3 pp, high confidence |
| HealthBench Professional | 59.6% | estimated ± 4.5 pp, medium confidence |
| HealthBench Professional (raw) | 66.0% | estimated ± 4.5 pp, high confidence |
| HLE-Verified | 54.3% | estimated ± 6.8 pp, low confidence |
| LABBench2 | 83.2% | estimated ± 2.4 pp, medium confidence |
| STEMBench | 2.4% | estimated ± 0.2 pp, high confidence |
| SimpleQA | 52.6% | estimated ± 4.3 pp, low confidence |
| Agents' Last Exam | 29.6% | measured |
| OpenHarmony Bench | 58.9% | estimated ± 2.4 pp, high confidence |
| BabyVision | 61.6% | estimated ± 11.4 pp, low confidence |
| MMVU | 77.4% | estimated ± 3.6 pp, medium confidence |
| MRCRv2 | 85.7% | estimated ± 11.3 pp, low confidence |
| MMLongBench-Doc | 61.4% | estimated ± 2.6 pp, low confidence |
| MMMU-Pro w/ Python | 81.9% | estimated ± 2.0 pp, high confidence |
| SWE-Atlas Codebase QnA | 63.7% | estimated ± 5.4 pp, low confidence |